Preprocessing: Added support to control features (with no transformat…

…ion), support to TSV as input, options to fill null with a median or a value

examples/quick_start/scripts/preproc/args_parsing.py

@@ -25,6 +25,7 @@ def build_arg_parser():
         help="",
     )
 
+    parser.add_argument("--control_features", default="", help="")
     parser.add_argument("--categorical_features", default="", help="")
     parser.add_argument("--continuous_features", default="", help="")
 
@@ -94,6 +95,7 @@ def parse_arguments():
     args = parser.parse_args()
 
     # Parsing list args
+    args.control_features = parse_list_arg(args.control_features)
     args.categorical_features = parse_list_arg(args.categorical_features)
     args.continuous_features = parse_list_arg(args.continuous_features)
 
@@ -110,7 +112,7 @@ def parse_arguments():
     if args.filter_query:
         args.filter_query = args.filter_query.replace('"', "")
 
-    if args.csv_sep.lower() == "<tab>":
+    if args.csv_sep.lower() == "[tab]":
         args.csv_sep = "\t"
 
     return args

examples/quick_start/scripts/preproc/preprocessing.py

@@ -164,12 +164,14 @@ def generate_nvt_workflow_features(self):
         args = self.args
         feats = dict()
 
+        for col in args.control_features:
+            feats[col] = [col]
         for col in args.categorical_features:
             feats[col] = [col] >> nvt_ops.Categorify()
         for col in args.continuous_features:
             feats[col] = [col]
             if args.continuous_features_fillna is not None:
-                if args.continuous_features_fillna.lower() == "<median>":
+                if args.continuous_features_fillna.lower() == "[median]":
                     feats[col] = feats[col] >> nvt_ops.FillMedian()
                 else:
                     feats[col] = feats[col] >> nvt_ops.FillMissing(args.continuous_features_fillna)